Anterior cruciate ligament reconstruction: middle-third bone/patella tendon/bone autograft
C.M. Fergusson
Department of Orthopaedic Surgery, Royal Berkshire and Battle Hospitals NHS Trust, Reading, U.K.
J.R.Coll.Surg.Edinb., 45, February 2000, 33-43

Keywords: anterior cruciate ligament, chronic instability, autograft, middle- third patella tendon

INTRODUCTION

Anterior cruciate ligament (ACL) rupture is a common, and initially often overlooked injury of the knee. Whether the true incidence of the injury is on the increase, or simply its professional recognition, is hard to say, but reconstruction for functional instability has developed rapidly over the last two decades. Not all ACL-deficient knees need stabilising, and unfortunately not all stabilised knees return to full function, so careful patient selection is very important in this group of patients. Once selected, adequate time must be spent in preoperative counselling to discuss the issues of the time implications, the critical importance of post-operative physiotherapy and rehabilitation, the risks of success and failure, as well as the complications, potential graft site morbidity and the alternative conservative approaches. These latter include proprioceptive physiotherapy and/or bracing for specific sporting activities. In support of this we provide a patient leaflet which includes details of the knee’s anatomy, surgical techniques, rehabilitation protocols, as well as expected outcomes and references.

The choice of graft for reconstruction is a more contentious issue. The middle-third patellar tendon graft remains for many the “gold standard” with the longest published follow-up results, but with a higher incidence of graft morbidity in the form of anterior knee pain and difficulty with kneeling than the four-strand hamstring reconstruction, its main rival. The use of hamstrings has seen a great increase in popularity in recent years with improvement in fixation techniques. There remains much debate about the two techniques as regards strength, stiffness, meniscal protection and anchor site problems - the main concern is the relative lack of long-term published data in the newer hamstring group. I reserve arthroscopic four-strand hamstring reconstruction, therefore, for patients before skeletal maturity, patients in a profession involving kneeling, patients with known anterior knee pain syndrome and revision cases, or those in whom there has been previous trauma to the patellar tendon.

Successful outcome of middle-third patellar tendon reconstruction relies on a scrupulous surgical technique, and closely supervised physiotherapy rehabilitation. The purpose of this article is obviously to attend to the former, but the physiotherapy rehabilitation is such an integral part of the success of the operation I have included our rehabilitation protocol at the end (Table 1).

PROCEDURE

Anaesthetic and Pre-Operative Set-Up

Subcutaneous low molecular weight heparin (2500m Fragmin^®) is given one hour pre-operatively. After induction of the general anaesthetic an IV bolus of 1.5 gm of Cefuroxime is given, after a test dose. A three-in-one block is inserted with a short bevel LA needle (30 ml of 0.5% plain bupivacaine for average sized adult), having identified the nerve with a nerve stimulator. A thorough examination under anaesthesia of knee stability and comparison with the opposite limb is carried out before the application of a tourniquet. Particular note is made of the degree of laxity, with gradings of the pivot shift, Lachmann’s and anterior drawer, as well as the presence of any secondary restraint failure of a specific nature (eg postero-lateral rotatory instability). A high thigh tourniquet is applied and a paper shut-off towel is taped carefully around the thigh distal to the tourniquet to prevent the skin preparation spirit from running back under the cuff, to prevent tourniquet burns (Figure 1).

Figure 1 Pre-operative positioning. Note the side support, the sandbag supporting the foot to maintain a 90° position and the drape stuck round the tourniquet to protect it from the spirit preparation

The patient is placed supine on the operating table with a side support at the level of the tourniquet cuff, and a sandbag at the level of the mid-calf. This allows the 90° position to be maintained without assistance and peri-operative adjustment of the sandbag will allow varying degrees of stable flexion to be achieved. The limb is exanguinated with a Rees-Davis tube at the last moment prior to skin preparation to reduce the tourniquet time, and to allow the intravenous antibiotics to achieve good bone levels. The leg is draped with disposable drapes, a stockinette to the thigh and an Ioban adhesive drape applied over a suitable-sized hole in the stockinette over the front of the knee.

Surgical Approach

A mid-line incision is made from the lower pole of the patella in a vertical and distal fashion to the patellar tubercle. This may need extension distally and proximally, but because of the mobility of the skin over the front of the knee, retraction at either end will give suitable access for the bone blocks to be resected. This can often be best achieved by an assistant’s hand resting on the patient’s skin and retracting it as a sleeve, rather than introducing retractors into the wound edge. A self-retaining retractor in the centre of the wound allows adequate exposure, and the vertical incision can be extended through the paratenon, which is then reflected back on either side to expose the full width of the tendon in its middle part. While it is necessary to see the tendon edges to assess a suitable width of graft, the dissection should not extend deep at the edges as there are visible vessels supplying the ligament and paratenon, which can be left unharmed.

Harvesting the Graft

While it is possible to measure the width of the graft with a ruler and make a calculation of the appropriate “third”, it is important to note that further application of the self-retaining retractor will widen the ligament and give a false width. It is also important to recognise that the central part of the tendon is significantly thicker than the outer part. A mathematical third of the width taken from the central piece will be more than a third of the strength. Tendons vary significantly between individual patients and it is easier to make an appropriate estimate in each case, and start by making a vertical incision in the ligament in its central portion approximately one third in, and extending it with a sharp side of the scalpel blade to the patella above and the tubercle below. It is important to allow the blade to find its own way through the planes between the fibres. Any cut fibre is a loss, not only to the graft, but also to the residual tendon. The ligament also narrows as it goes distally, and as it does, so must the graft. Being aware of this subtle but important anatomical point allows a freehand graft to be safely taken without dividing the fibres. The use of a template or even a parallel cutting jig is discouraged for this reason, but if one wishes, a sterile pen may be used to draw out the cuts beforehand. Once the graft has been outlined by these incisions they can simply be extended onto the periosteum of the patella and the tibial tubercle in a parallel fashion with the knife blade, for approximately 30 mm from the insertion at either end. The graft shape is completed by bringing the periosteal incisions to a point.

Once outlined, drill holes are made at both ends through the cortex (Figure 2).

Figure 2 After the skin incision and the middle-third has been marked out, patellar drill hole is being made. Note the hand supporting the instrument as a depth stop

One should be approximately 5 mm from the tip of the proposed graft, and a second hole drilled 1 cm further back. It is possible to drill these holes once the graft has been taken out, but it is more difficult to place them accurately. Once the drill holes have been made, the cortical bone at the patella and tuberosity site is divided with a power saw angling the blade inwards towards the midline (Figure 3). It is easy to feel when the cortex is completely cut at both sites, as the blade falls easily into the underlying cancellous zone.

Figure 3 Patellar bone block being cut with power tool. Once again the drill being supported to prevent over-penetration

At the top and bottom of the graft the saw blade can be angled along the pointed segment to complete the cortical cut. The purpose of this is to reduce the transverse cut at either site. It reduces the risk of any residual surface fibres being damaged by the blade, and also at the patella site. If there were a fracture, this would almost certainly propagate in a vertical fashion and be far less damaging to the extensor mechanism and easier to attend to. Once the cortical cuts have been made a small osteotome is used to complete the graft removal. This is not intended to cut the cortical bone, but simply to rock the graft side to side on its underlying cancellous pedicle, which will eventually break through, working the osteotome from one cut to the other. Thus, if after the first gentle tap with your toffee hammer, the graft does not move, the cortical cuts need further attention with the saw. By gently freeing the graft in its cancellous bed at each end it will quickly come free and cutting a few remaining fibres from the fat pad will finally free the graft.

Graft Preparation

The sharp bone ends can be trimmed back and the cancellous bone rounded off with the bone nibblers, and all bone chips should be saved for subsequent re-introduction to the donor site. A gallipot, half-filled with normal saline, is an appropriate receptacle. The bone ends are shaped until they comfortably fit inside the sizing tubes (Figure 4).

Figure 4 Sizing the bone blocks with the “Acufex” tubes

Normally at this stage one end will measure 9 mm and the other 10 mm, and the bone blocks will be approximately 30 mm in length. The soft tissue and any adherent fat is cleared away from the undersurface of the graft , so that visualisation as it is passing through the joint is made easier. A size 2 Ethibond^® suture is passed through the smaller of the two graft ends at the tip drill hole, identifying it as the femoral end, and a stainless steel wire is put through the larger of the two ends. A single turn in the wire emerging on the cancellous side of the distal bone graft will identify this surface for alignment in the tibial tunnel (Figure 5).

Figure 5 To show the wire holding the larger of the two bone blocks and the Ethibond® suture holding the smaller leading end

Arthroscopy

I perform an arthroscopy through the original incision to reduce further skin wounds; the point of scope insertion can be identified antero-laterally next to the tendon avoiding damaging it (Figure 6).

Figure 6 The use of the anterior incision to gain access to the antero-lateral portal site without creating a second skin incision

The central portal through the fat pad (i.e. the donor site) can be used for the hook, and for most meniscal surgery at this stage if necessary, but a further antero-medial portal can be used if indicated. A careful diagnostic arthroscopy is completed to identify any meniscal pathology, and attend to it with repair resection or trimming as indicated (Figure 7).

Figure 7 Meniscal surgery arthroscopically with the instruments inserted through the central portion of the incision

By the same token articular surface damage is noted and with this approach it is possible to do mozaicplasty allografting, if indicated, with a combination of arthroscopic and mini-open technique.

Mini-Arthrotomy

A small section of the fat pad can now be excised to gain access to the anterior part of the notch and the ACL stump. A small self-retaining retractor will expose this area, although it will be necessary to extend the knee somewhat and carefully adjust the light to view the notch. Dissection with an 11-blade on a long handle and pituitary rongeurs will now give excellent view of the notch and its contents. In most cases a small lateral notchplasty and roofplasty, taking a few millimetres of the joint margin at most, will enhance the view, prevent impingement and remove any small osteophytes or constitutional narrowing which may compromise the graft function. In many cases it is not necessary to do more than simply tidy up the edges of the notch itself. The contents of the notch are dissected to reveal the sharp vertical border of the lateral side of the posterior cruciate ligament (PCL), which defines the medial border of the dissection. The notch is emptied of scar tissue, old ACL remnants and fatty tissue until it is possible to see the dark recess at the back of the notch (Figure 8), and accurately probe the over-the-top position with the arthro-scopic hook. (The over-the-top position is at the back of the roof of the intercondylar notch, off-centre - 11 o’clock on the right, 1 o’clock on the left - where further dissection would take you over-the-top of the lateral femoral condyle).

Figure 8 A view through the small arthrotomy between the outer thirds of the patellar tendon showing the prepared lateral wall with a small notchplasty

This must be felt as well as visualised, as anterior placement of the femoral position is a common cause of failure. The Volkmann’s curette is a very useful tool for the final preparation of this part of the procedure. Only once the over-the-top position can be clearly identified and felt should one proceed.

Tibial Tunnel

With sharp and blunt periosteal dissection a small area of the upper medial tibia can be exposed next to the lower part of the wound, medial to the tubercle. This should be sufficiently distal to allow a good length of tunnel to get sound fixation at the distal end, which is the most critical for strength, but the pes ansurinus insertion should not be encroached upon. A heavy guide wire is then drilled from the tibial side into the notch and should enter the knee centrally, and just in front of the PCL. This is done under direct vision so it is safe to do it without a jig if one is confident in one’s aiming. The mistake is to bring this guide wire too far forward and suffer the problems of impingement of the graft in full extension. It is practically impossible to put it too far back, as long as it is in view in front of the PCL.

Having stressed that the positioning is critical, to some extent the reamer, which is passed up the guide wire subsequently, has a mind of its own and centres itself between the tibial spines and tends to break out anteriorly because of the oblique angle of the reaming. Since the distal reamed hole will be the same size as the larger of the graft ends (ie 10 or 11 mm) and this is the width of the tibial eminence at this site, as long as the guide wire comes out between the spines in the mid-line, the reamed hole will be soundly placed. As the reamer is being used, the bone fragments emerging can be collected with a Volkmann’s spoon and added to the half-filled saline pot for later return to the bone defects. The posterior part of the tibial tunnel is chamfered with the angled rasp so there are no sharp edges posteriorly, and any soft tissue “trap door” should be excised from the opening into the knee joint. Any potential impingement can be checked for now, by placing a long thick instrument (eg a rongeur) up the tibial tunnel and extending the knee; any ‘capture’ of the instrument by the notch roof will be easily detected.

Femoral Tunnel

A right angle offset jig is passed up the tibial tunnel, run along the roof of the femoral notch until it is felt to click over the back of the over-the-top position (Figure 9).

Figure 9 The right angle jig in place up the tibial tunnel ready to accept the guide wire through the femoral insertion point

This is as easily felt as seen, but confirmation by the eye is important and, if necessary, with the arthroscope which can be used in the dry state in this mini-open technique. The knee can be very gently flexed until the jig is gripped in place.

The distal tongue of the jig is offset 6 mm back from the central hole, thus allowing the user to ream a hole with a diameter of 10mm and still leaving a small distal wall. The guide wire is then driven up through this jig, out through the top of the femur and skin and rescued through the stockinette, and a reamer of the size to match the smaller bone block is drilled to the same depth as the graft is long. The depth of the reamer can be assessed with the arthroscope most accurately. The bone fragments are rescued with the rongeurs and added to the gallipot, and the arthroscope with the water now running can be used to flush out any remaining fragments from the tunnels, both tibial and femoral, and to check the posterior wall of the femoral drill hole is suitably sound for an interference fit screw (Figure 10).

Figure 10 The use of the arthroscope through the notch to check the femoral drill hole and wash out the debris

Placement of the Graft

The Ethibond^® sutures are now threaded through the guide wire like a needle and with a pair of pliers the guide wire is pulled out from the femur, and the Ethibond^® sutures held with haemostats. The graft is properly orientated so that the two cancellous surfaces are facing forward; by pulling on the Ethibond^® sutures the graft is gently introduced up the tibial tunnel, across the knee and into the femoral tunnel. At its entry into the femoral tunnel it will inevitably get stuck. Finger pressure applied to the base of the graft pushing posteriorly will tip its leading edge into the femoral hole, and slight flexion/extension of the knee will see it slide gently into the position. At the same time as it is negotiating the proximal end, the distal end will be entering the tibial tunnel (Figure 11); it commonly gets stuck at this point and is a cause of frustration.

Figure 11 Insertion of the graft into the tibial tunnel. A small turn on the wire can be seen which identifies the cancellous side of the bone block to allow correct orientation for screwing

If the tibial tunnel is a tight fit it can be pushed firmly in first, to allow some slack along the length of the graft. Although the graft tends to rotate somewhat it can be positioned easily, either with a finger tip or a pituitary rongeurs, so that the longitudinal saw cut is facing anterior and is clearly visible. The graft is pulled in until only 1 cm remains showing. At this point a flexible guide wire can be positioned within the groove, and by angling it with finger pressure, can be pushed up along the front surface of the graft. On most occasions, the accurate placement of this will be confirmed as it comes out of the anterior femoral hole. Using this groove technique avoids a typical complication of the graft being crossed by the screw, and the fixation threatened by inadequate interference. Once the guide wire is soundly in place anterior to the cancellous surface, the graft is seated by a final pull on the Ethibond^® suture. The bone should just disappear out of site.

Inserting the Femoral Screw

The common problem to avoid here is bending the guide wire by trying to insert the screw on the wire, which is curved anteriorly round the corner. To avoid this, when the wire is in place the knee is hyperflexed. As you flex the knee you can watch the guide wire straighten out and eventually lift off the fat pad at the front of the mini-arthrotomy (Figure 12).

Figure 12 On full hyper-flexion the guide wire is seen to lift free of the fat pad confirming that it is in a straight line and can therefore safely accept the screw without kinking the wire

At this point, the wire can be assumed to be horizontal and straight. Insertion of the screw along this track will now be straightforward. A 7.5 mm x 25 mm screw is used, and advanced up the screw towards your hand pulling on the Ethibond^® suture. This confirms the alignment.

At the same time an assistant takes up the tension gently on the wires distally, so that the screw does not bind up the ligament and keeps it clear of the advancing screw edges. A firm push is required to seat the screw to start with, but after two or three turns it will start to run up the guide wire easily. After a few more turns, the guide wire can be removed and the screw is driven home, finally clicking as it bites into sound bone, fixing the graft in place. The screw is inserted until it is just deep to the bone tunnel opening, and this can be confirmed with the arthroscope. The 7.5 mm screw diameter is used in order to avoid a blowout of the posterior wall, but in the unlikely event of the graft coming free on tension at this stage then a larger screw may be required. This has only occurred on very rare occasions.

Tensioning of the Graft and Fixation of the Tibial End

Firm tension downwards on the wire will confirm solid fixation at the femoral end. By cycling the knee, maintaining the tension on the graft at the tibial end, the surgeon will confirm isometricity of the placement and take out any slack in the graft itself. As long as the graft has been fixed suitably posteriorly, as above, the graft will be isometric through the main range of movement, slightly paying out in full hyperflexion and slightly going in at the hyperextension range of movement when the tibia has pivoted forward, which of course is the movement we are trying to overcome. By fixation just 30° off extension before the pivot point is achieved, allows the knee to be stabilised without being captive. A firm pull on a haemostat clipped on the wires distally will allow tension in the graft, and the right degree of slight flexion. A guide wire is passed up anteriorly and is seen to enter the joint and lie in front of the graft to be sure that the screw fixation is anterior. Usually, the orientation of the bone block is clear, but the single turn on the cancellous side of the wire will ensure that there has not been 180° rotation. It is important that the screw threads do not run up the non-cancellous surface and damage the graft. A 9 x 30 mm interference screw is used at this site, as the bone is softer there is no significant blowout risk, and the fixation is even more important. This is because, when the graft is functioning properly near extension, the pull on the graft is vertically along the line of it, whereas that on the femoral site is at almost 90° and, therefore, has a much greater mechanical advantage. Nevertheless, with this technique there is normally a satisfying click (or several), as the screw is driven home up the guide wire and entered no further than the end of the graft itself which is approximately the screw length. Any residual protruding tag of bone can be trimmed so that the tibial surface is flush. The wire and Ethibond^® suture can now be pulled out.

Final Checks and Closure

It is essential that the reconstructed leg has an unrestricted range of extension. It must also be stable and there must be no impingement on the notch anteriorly. This latter can be checked with the arthroscope as the leg comes into extension (Figure 13).

Figure 13 The leg in extension and the scope being used to ensure no impingement of the graft anteriorly

This technique will control the Lachmann’s and pivot shift completely at the end of the operation. A slight anterior drawer can still be identified in some patients and this is not of functional consequence, but simply reflects the limitations of the technique and the rather posterior positioning of the graft .

The mini-arthrotomy is closed with absorbable Dexon^® sutures to the fat pad, the bone chips and reaming fragments are retrieved by pouring off the saline from the gallipot, and the donor sites can be packed (Figure 14).

Figure 14 The retrieved reamings are packed into the donor area in the patella

There is normally more than enough to adequately fill these to the surface. They can be held in place by a layer of sutures in the periosteum, and the paratenon is closed but no sutures are put through the remaining graft itself. There was some concern about the development of patellar infera by over-suturing the graft site. With the modern techniques of rehabilitation, however, this has not been seen, and closure of the paratenon is sound and safe, even though it may appose the tendon edges in the central portion. Clearly, sutures through the tendon itself will do nothing more than damage the longitudinal fibres. The superficial fascia is closed with absorbable sutures and a sub-cuticular Monocryl^® suture is used with steristrip support to complete the closure. An adhesive Mepore^® dressing is used and a light wool and crepe applied. I do not drain the wound. The leg is placed in a lightweight cricket pad splint in extension, which protects the limb from passive movement until muscular control returns.

Post-operatively, the limb is elevated, toe observations started, analgesia continued to support the three-in-one block, antibiotics are given with two further boluses of 750 mg. Cefuroxime at 8-hourly intervals. Anticoagulant therapy continues with subcutaneous low molecular weight Heparin until discharge on day 3 (2 500 m Fragmin 12 hours after 1st dose, daily doses of 5 000 m). Radiographs are taken post-operatively to ensure sound fixation and correct positioning, and are a useful source of auditing the techniques described. Early rehabilitation is devoted to achieving passive hyperextension and maintaining analgesia (opiates, Paracetamol and nonsteroidal anti-inflammatory drugs) after 24 hours) and reducing swelling with ice packs and cold therapy if necessary. CPM is not used and flexion return comes naturally if hyperextension regimens are maintained. The rehabilitation continues on the lines indicated below, with particular reference to early kneeling regimes to reduce scar hypersensitivity and donor site morbidity.

TROUBLESHOOTING

This procedure is technically challenging and there are minor and major problems that can occur. These are some that I have either experienced or have been asked about, and have been reported in the literature:

Fracture or Pull-Off of One of the Bony Ends of the Graft

In this case, the graft can be reversed so that the remaining bone end is used in the femoral tunnel, the free end whip-stitched and held in the tibial tunnel with interference fit screw, and reinforced by bringing the whip-stitch down and round a screw and washer at the tibial surface. If a small part of the tip of the bone graft fractures, using the secondary hole and trimming the graft appropriately is all that is required.

Blowout of the Femoral Side

If this occurs, a small lateral femoral condylar incision can be made onto the one surface and, using an appropriate jig system, a separate hole can be drilled down to the isometric point. The femoral bone plug can then be pulled back up into this and held with an interference screw from the femoral side.

Guide Wire Breakage

If the small screw guide wires get bent on insertion the tip can break off. I have had one case in which I left the guide wire tip in situ. The functional result was sound without any complications and I did not, therefore, retrieve it.

The Graft is Too Long

The original description of the operation with the tibial jig prevents this happening. The technique I have described above short cuts the system, as I make a tibial hole fairly low down and I have not faced a problem of an over-protuberant graft. Nevertheless, if it does occur, if the graft is no more than 1.5 cm protruding, much of this will be taken up by impaction as the interference screw is inserted. As long as the bone block is long enough, good grip will be achieved and a surprisingly little amount of the bone end will need to be trimmed flush with the tibial surface. If the graft is too long(exceptional circumstances), a tibial gutter could be cut below the tunnel and small fragment screw or screws used to hold it in place. If this were not practical, leaving an implantable wire around the graft over a transversely placed screw and washer would be an alternative.

Dropping the Graft on the Floor

Apart from stressing the obvious precautionary measures (preparing the graft in an appropriate position and pre-drilling the holes before cutting the graft), this is always a potential hazard and although I have never seen it, it has been reported. Apparently, washing the graft copiously and soaking it in Bacitracin^® has allowed it to be implanted safely without risk of sepsis. With the modern techniques of hamstring surgery available, however, if it now occurred I would change and use the hamstring from the same site, rather than risk deep sepsis in the knee.

GOALS OF SURGERY

In general, the aim must be to leave the knee stable and allow early rehabilitation. Ultimately, I would favour going for rehabilitation, even in the possible presence of a doubtful fixation, rather than to try and protect it with all the attendant risks of cast, fixed braces and so forth, that were once used. If all else fails the ipsilateral hamstring would be an appropriate alternative. Only in revision cases with appropriate preoperative counselling, would I use the contra-lateral hamstring.

Table 1: Rehabilitation regime for anterior cruciate ligament reconstruction patients. This regime applies to most cases but variation will occur depending on an individual’s severity of injury and other factors. Regime valid at the Department of Orthopaedic Surgery, Royal Berkshire Hospital, until March 2000